Skip to content

Accounting for sample imbalance in k-mean #1

Description

@william-denault

Dear Yiqun Chen,

Thank you for releasing this very nice package. I am interested in using your package for analyzing some spatial transcriptomics data. I have two main questions:

First, the amount of data (i.e., the number of cells observed) varies substantially across individuals. Therefore I would like that each individual contribute 'equally' to the inference results. Therefore, I would have liked to 'reweight' my observation in a way that is inversely proportional to the number of cells they "provide" (what you would in the lm function using the weight argument). Is there any weight to reweight my observation within the kmeans_estimation function?

Second, if I understand correctly, your approach is testing the difference in the overall mean vector between clusters. However, I am mostly interested in the difference between specific components a mean vectors. In other words, is there a way to test which of the entry drive the actual different between the two mean vectors of two clusters while controlling the Type I error?

Thank you in advance for your answer.

Best regards,
William Denault

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions